Incremental printing of symbolic information – Ink jet – Ejector mechanism
Reexamination Certificate
1999-08-18
2001-07-17
Nguyen, Thinh (Department: 2861)
Incremental printing of symbolic information
Ink jet
Ejector mechanism
C347S019000, C358S001900
Reexamination Certificate
active
06260948
ABSTRACT:
FIELD OF THE INVENTION
This invention relates generally to machines and procedures for printing text or images (whether photo-like pictorial images or commercial-style graphics such as charts, graphs, color bands behind text, etc.) from individual ink spots created in a two-dimensional pixel array on a printing medium such as paper, transparency stock, or other glossy media. The invention relates more particularly to preliminary procedures that prepare data assemblages for later use in guiding the operation of an inkjet machine and method, and most particularly to formation of lookup tables that enable such a machine and method to select inkdrop combinations quickly, efficiently and ideally. The invention is also applicable, however, to other methodologies such as hot-wax transfer or xerographic printing. The invention was made for use in printing by error-diffusion techniques.
BACKGROUND OF THE INVENTION
1. Possibly Relevant Patents
A search of the patent literature returned the following U.S. patents, as well as the coowned patents mentioned earlier. U.S. Pat. Nos.
5,331,440 Kita
5,425,134 Ishida
5,719,956 Ogatsu
5,737,453 Ostromoukhov
5,739,917 Shu
5,742,405 Spaulding
5,805,178 Silverbrook
5,809,181 Metcalfe
5,857,063 Poe.
Of these, as will be seen the most relevant appears to be U.S. Pat. No. 5,739,917, issued to Joseph Shu of Seiko Epson.
2. Approximating a Color Continuum with a Limited Number of Discrete Colors
Photograph-like pictures are generally displayed on computer screens using “twenty-four-bit color”—a phrase which refers to eight bits for each of the additive primary colors red, green and blue. For each primary color the eight bits provide 2
8
=256 possible levels, ranging from none of the specified color to full saturation of that color.
Of course not all colors are primaries, but the computer screen can show combinations of any level of each of the three primaries. Therefore the number of possible colors that can be displayed in any single pixel is 256
3
, which comes to nearly seventeen million discrete colors.
Unfortunately most incremental printers—inkjet, or xerographic printers, for example—have a much smaller set of actually printable colorants. The simplest and best known of these devices is binary and usually provides the three subtractive primary colors cyan, magenta and yellow—plus black. The number of discrete colors that can be printed with such a unit is therefore only 2
4
=16 colors within a single pixel.
Some more-modern devices, however, instead have two different dilutions of some of the colorants—usually of magenta and of cyan, and sometimes of black. Furthermore, these devices may be able to provide varying numbers of inkdrops (or other quanta) of the colorants, for instance from zero to four drops of each of the different colorants in their different dilutions.
As a result, the number of discrete colorant combinations that can be produced within any single pixel may be, say, into the thousands. Even these numbers are obviously far smaller than seventeen million. As a practical matter furthermore, many of these combinations are very close to one another and hence essentially redundant—so there are not really as many discrete colors as there are of colorant combinations. Moreover many of the combinations are best forbidden because they would deposit too many drops of ink (too much liquid) in a single pixel.
Consequently, as will be seen later, the number of practically useful discrete colorant combinations may be roughly one hundred twenty-five. How can a 16-color or 125-color machine (or even a thousand-color machine) make colors that look like the colors in an original picture, if the original was able to use any of seventeen million colors?
What is usually done is to trade off some spatial resolution (very small pixels) for “color space” resolution. In other words, some of the printer's capability to produce extremely fine detail is sacrificed, and the system averages the available colors over some relatively large number of pixels.
The system accepts coarser resolution within the two-dimensional positional space of the image to obtain finer resolution within abstract color space. In this way, much finer color gradations are obtained.
The number of pixels used in the averaging process determines how close to 17 million colors the printer can get. One way in which this tradeoff is done is called “dithering”.
Dithering is not closely relevant to the present invention. In the dithering approach a fixed, well-defined, usually rectangular cell of pixels (generally between two and thirty-two pixels) is used to produce a kind of color averaging within the cell. Although dithering works well for commercial graphics and other images that contain extended fields of uniform color, it has a tendency to generate within the image spurious visible patterning that is not usually acceptable for photograph-like images.
For photos, therefore, most workers prefer a different approach called “error diffusion”. This system evaluates the original twenty-four-bit input color in a specific pixel to see what color available in the machine (for instance, which one of the one hundred twenty-five discrete colors that a machine can make in a single pixel) is closest to that input color.
Then the printer selects to print, and actually prints, that closest input color in the specific pixel—but it also evaluates the error in the printed color. This error is then distributed to several nearby pixels that have not yet been processed.
When the system later reaches one of those pixels, for processing, it adds to the input color in that pixel the previously allocated error from earlier-processed pixels—before making the assessment and printing decision mentioned above. In this way error is continuously propagated from pixel to pixel, and so is diffused, analogously to some component liquid in a continuous-dilution process in a liquid stream.
3. Basics of Underlying Device-State System
The present inventors' above-mentioned prior patent document advances the art of multiple-ink error diffusion by defining so-called “device states”. This is where the one hundred twenty-five discrete colors come from—they are carefully precalculated and refined to provide an ideal “palette” of basic colors from which good approximations to all colors can be error-diffusion generated.
The device states, or palette colors, are worked out so that they not only provide an excellent source of colors for use in combinations but also are optimized in terms of the amount of liquid going into each pixel. The system of the related patent application operates very quickly and efficiently, based upon a lookup-table (LUT) approach in which not only the target device state but also the corresponding error is found in the table, thereby saving much computational time.
Thus the LUT has several thousand “major entries” from which to funnel down to the one hundred twenty-five available device states. The related patent application also introduces other procedural features that are significant overall but not important to the now-desired patent search.
4. The Printing of Grays; Black Replacement and Undercolor Substitution
An important and difficult aspect of color printing by almost any methodology is the selection of colorants for reproducing gray and near-gray constituents of colors, to produce best image quality—and particularly quality in highlight areas. First, grays and near grays may be regarded as dilutions of “black”.
There are two principal ways to print black, whether in incremental printing or in older-fashioned printing-press systems: “single color” black, using real black ink, and “process black” which is produced as a common quantity of the three subtractive primaries cyan, magenta and yellow.
It is well known that in many situations it is desirable to substitute real black ink for a common fraction of the three subtractive primaries that may happen to be present in a particular desired color. This is a desirable thing to do in midtone regions of an image and also in
Bockman Francis E
Li Guo
Hewlett--Packard Company
Nguyen Thinh
LandOfFree
Reducing granularity in highlights and yellow fields by... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Reducing granularity in highlights and yellow fields by..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Reducing granularity in highlights and yellow fields by... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2465218